For Every Public Policy There are Corresponding Private Interests

I learned a long time ago, as a student of public policy and then as a journalist, that “for every public policy there are corresponding private interests.”

So arguments about whether and how to regulate artificial intelligence in the context of content businesses always will be a combination of abstract public values, impact on culture, fairness, content quality or art and perceived personal economic interest.

Every major technological shift affecting content industries has altered

• who creates value

• who captures income

• whose social status changes. 

Indeed, much of the economic value in content industries rests on scarcity, and AI threatens to create abundance. That might be a favorable outcome for content consumers, but might harm professional content producers. 

Traditional source of scarcity	Effect of AI
Skilled illustration	AI greatly expands supply
Copywriting	Near-zero marginal production cost
Translation	Instant multilingual capability
Stock photography	Synthetic images substitute for many uses
Voice acting	Synthetic voices compete in many applications
Video production	Increasing automation reduces labor inputs
Software documentation	AI drafts much routine material
Marketing content	Mass personalization becomes inexpensive

Whenever scarcity declines, prices usually follow. So the content industry advocates concern about AI "ethics” also are about protecting existing economic rents. 

That isn’t unusual. All professional associations, licensing requirements and unions, whatever their stated purpose (“safety,” often), are also about protecting economic rents.

Public policy concern	Corresponding private interest
Copyright protection	Licensing revenues
Artist consent	Control over monetization
Transparency	Ability to distinguish human work in the market
Watermarking	Preserve premium pricing for human-created work
Fair compensation	Maintain existing wage levels
Quality concerns	Preserve professional gatekeeping
Educational concerns	Preserve demand for traditional instruction
Safety regulation	Increase barriers to entry favoring incumbents
Cultural preservation	Preserve existing creative institutions

We can cite many content industry examples.

Technology	Incumbents defending existing value	Public argument	Private interest
Printing press	Scribes	Accuracy, religious authority	Preserve copying profession
Photography	Portrait painters	Artistic standards	Maintain commissions
Recorded music	Live performers	Artistic integrity	Preserve performance income
Radio	Newspapers	Media concentration	Advertising revenues
Television	Movie theaters	Cultural effects	Box office
Digital photography	Film manufacturers	Image quality	Film sales
MP3 files	Record labels	Copyright	Music distribution revenues
Streaming	Cable operators	Local programming	Subscription economics
Generative AI	Writers, artists, actors, publishers	Copyright, authenticity, quality	Employment, licensing, bargaining power

Also, AI threatens not only earnings but also professional identity, as creative professions provide:

• expertise

• prestige

• cultural influence

• reputation

• gatekeeping authority

• community standing. 

If AI enables non-experts to produce acceptable commercial work, professionals may lose status even before they lose substantial income.

The broader lesson from economic history is that technological debates are rarely contests between "public good" and "private greed." 

Instead, all public policies have corresponding private interests.

AI raises authentic questions about authorship, consent, cultural diversity, and market power. 

At the same time, it redistributes income, bargaining power, and professional status across the content ecosystem. 

That isn’t to deny the legitimacy of the issues raised. But neither does it make sense to deny the private financial interests also at stake. 

Good Outcomes Matter More Than Good Intentions

California voters will decide the fate of Proposition 40, a new wealth tax, in November 2026. The initiative would enact a one-time tax of five percent on the accumulated wealth of taxpayers and trusts with covered assets valued over $1 billion. 

As popular as such “soak the rich” policies might be in some quarters, governments imposing such wealth taxes have found highly mixed returns from the policies, the Organization for Economic Co-operation and Development says. 

Proponents of such taxes might tout the equity benefits. Critics might point out that high-net-worth individuals can, and do, simply move to avoid the taxes. 

But there are other issues, including the reality that wealth taxes often raise less revenue than expected, while imposing disproportionate economic and administrative costs. 

And such laws, where they have been imposed, are being repealed. In 1990, 12 OECD countries had such taxes. By 2017 there were just four OECD countries that continued to do so. 

The OECD report  notes that many countries repealed wealth taxes because they:

• generated relatively little revenue

• were expensive to administer

• encouraged avoidance

• were perceived as economically inefficient

• often failed to achieve redistribution objectives as intended.

So migration of taxpayers is one of several objections to such taxes, which increasingly are possible when assets are internationally diversified, which increasingly is the case for such high-net-worth persons.

Migration was rarely the official reason cited by governments which repealed such taxes. The OECD notes that wealth taxes typically produced surprisingly small amounts of revenue compared with expectations.

OECD does not say such policies can never work. “For instance, a net wealth tax may have more limited distortive effects and be more justified as a way to enhance progressivity in countries where the taxation of personal capital income is comparatively low,” the report says. 

“Overall, the report concludes that from both an efficiency and equity perspective, there are limited arguments for having a net wealth tax in addition to broad-based personal capital income taxes and well-designed inheritance and gift taxes,” the authors note.

Considerations of equity benefits aside, one of the strongest criticisms of such taxes is their modest fiscal yield, as they generate small returns:

• usually well under one percent of gross domestic product

• generally a small share of total tax revenue;

• often lower than expected because of exemptions, avoidance, valuation challenges, and migration.

The OECD repeatedly notes that low revenue was a major reason countries abandoned these taxes. In other words, the policies generally do not work. 

But the persistence of wealth taxes in countries such as Switzerland, Norway and Spain also shows that outcomes depend heavily on policy design, tax rates, exemptions, and the broader tax system, rather than on the mere existence of a wealth tax, the report suggests. 

It is unclear whether the California measure would produce meaningful revenue or not, but the key point is that, with all public policy, having good intentions is one thing. 

But that matters less than actual good outcomes. Some of us are likely doubtful the measure’s actual stated goals can be achieved, in practice. 

So it is in the category of actions we might say are mostly theatrical and symbolic. To the extent the measure would encourage migration out of the state, which would tend to lower tax receipts, the measure might even be counter productive. 

Enterprise AI Market Still Has No Settled Market Share Structure

After analyzing more than 100 artificial intelligence products used by enterprises from 2022 to 2026, researchers at Okta find adoption patterns you might well expect:

• Anthropic’s growth leads all others

• The market still has not settled into a stable pattern

• Multi-vendor is a common approach

• Use of specialized products is significant. 

source: Okta

The analysis suggests a widespread multi-vendor adoption pattern as well. 

source: Okta

AI Capex to Exceed $1.2 Trillion in 2027 and 2028?

Is the growing cost of Nvidia-based compute a business moat or a liability? And in either case, which contestants benefit or suffer? 

A new Morgan Stanley analysis suggests the cost of Nvidia's GB200 systems now cost about $35 billion per gigawatt (GW) of computing capacity, up 16 percent from prior estimates. 

GB300 clusters rise to $39 billion per GW, while Vera Rubin-based systems jump nearly 20% to $49 billion per GW.

Those estimates include networking equipment, storage, liquid cooling systems, electrical infrastructure, and power.

A single GW can power roughly 700,000 to 1 million U.S. homes. 

So are high costs a business moat or a sign of dangerous costs, or perhaps both?

Morgan Stanley also projected that the combined capital expenditure of the five largest AI-driven firms (Microsoft, Google, Amazon, Meta, and SpaceX) will reach approximately $1.2 trillion and $1.4 trillion in 2027 and 2028, respectively. 

By 2028, available computing capacity is expected to grow from around 30 GW in 2025 to nearly 120 GW, a fourfold increase.

Morgan Stanley expects the combined available computing capacity  of the five major hyperscalers to reach nearly 120 GW by 2028, a fourfold increase from approximately 30 GW in 2025. 

AWS will have the largest capacity in 2028 at 35 GW, followed by Google at 31 GW. Meta’s capacity is projected to grow from approximately 3.5 GW at the end of 2025 to 14 GW in 2027 and 21 GW in 2028.

But the AI compute arms race is shifting from "how much you build" to "how much you sell,” Morgan Stanley suggests. Who can convert compute into monetizable revenue streams?

AI Copyright Balance Will Emerge: All Prior Content Technology Innovations Have Done So

Copyright is a tricky business, and has been since the advent of digital media. But artificial intelligence likely will cause a rethinking or adaptation of copyright, in part because it is getting harder to distinguish between a human author’s particular formulation of an idea and an AI-generated alternative. 

Traditionally, copyright has been designed to encourage innovation by providing creators with limited monopolies over their work, protecting the particular expressions of ideas, but not the ideas themselves.

So copyright protects an author’s expression of an idea, not ideas, facts, or styles themselves. 

Abundance, instead of scarcity, partly explains why that is the case.

Copyright traditionally assumed scarcity and control over physical copies (books, records, films). All that became more challenging when digital distribution; the internet; user-generated content and easy remixing of content replaced scarcity and distribution cost:

• Digital formats (MP3s, JPEGs, PDFs) allowed lossless copying at near-zero cost, unlike analog media. Napster (late 1990s) and peer-to-peer sharing exemplified mass infringement

• Global, instantaneous sharing by websites, torrents, and streaming bypassed traditional gatekeepers

• Social media, YouTube and many other sites blurred lines between consumers and creators, increasing derivative works, remixes, and mashups

Era	Key Challenges	Fair Use Evolution	Legislative,  Policy Responses
Pre-Digital (pre-1990s)	Physical copying limits; analog scarcity	Narrow: criticism, parody, education	1976 Copyright Act  codifies fair use & rights
Digital Media and Internet (1990s-2010s)	Perfect digital copies, P2P sharing, online distribution	Expanded for search/indexing (Google Books), software interoperability (Google v. Oracle)	DMCA (1998): safe harbors, anti-circumvention; longer terms
AI/Generative (2020s-)	Mass scraping for training, output substitution, authorship	Split rulings: often transformative for training (Anthropic, Meta) but market harm weighs heavily; fact-specific	Ongoing reports (US Copyright Office, UK, EU); licensing debates; deepfake proposals

Generative AI arguably complicated matters further:

• Model training is based on use of massive datasets, often scraped from the internet, raising questions of reproduction rights. AI companies argue it's necessary for learning patterns/ideas (not protected expression); creators call it systemic infringement.

• AI can generate content mimicking styles, potentially causing "market dilution" as a new issue

• AI outputs generally lack human authorship for copyright, but human-prompted or edited works are gray areas

• Training data is often non-transparent and opt-out mechanisms are impractical at scale

• U.S. courts show diverging fair use rulings; reports from U.S. Copyright Office, UK, EU Parliament highlight needs for licensing, transparency, or new frameworks (e.g., compulsory licensing debates).

The main policy choices are whether to create new rules for training data, how to treat AI-generated outputs and whether to add special protections for likeness, voice, or style cloning.

The core tradeoffs, as always, are between innovation and creator compensation. Broader licensing can raise costs for AI developers and small firms, but it may also reduce litigation and give rights holders a clearer market. 

Looser rules can speed model development, but they may weaken incentives for human creators.

For now, the U.S. Copyright Office says existing copyright principles are flexible enough to handle AI. 

As with prior content industry technology changes, some balance will emerge that compensates content owners and also allows innovation. 

AI Productivity is Notoriously Hard to Measure

Observers say artificial intelligence often changes “how work is done” or “how well work is done” (quality improvements) rather than just “how fast it is done,” leading to outcomes that are difficult to capture in traditional productivity statistics. 

That is particularly true for intangible products such as software, e-books, downloadable music, mobile applications, healthcare consultations, financial advice, legal services, streaming subscriptions, web hosting or any other product that is experiential (haircuts or live concerts). 

Product quality changes, called “hedonic,” are particularly hard to quantify in these cases. Among the classic examples are personal computers that, over time, incorporate faster processors, more memory, better user interfaces, displays or audio, but without a price increase. 

Smartphones might add premium materials, for example.

The point is that much of AI’s value is qualitative: improved decision-making, better user experiences, or reduced risk in complex processes (like drug discovery) that will not always show up as an immediate increase in volume-based output.

And all that is hard to measure. 

Proxy Metric	What it Measures	Limitation
Task Completion Time	How much faster a specific, defined task is finished with AI.	Ignores quality variance and "rework" time (verification).
User/Adoption Rates	The percentage of the workforce actively using AI tools.	Does not measure value or net efficiency gains.
Resource Optimization	Reduction in compute or operational costs for a given output.	Can hide negative impacts on employee skill formation.
User Satisfaction	Improved quality of output or speed as perceived by the customer.	Subjective and may not correlate to bottom-line profitability.
Error/Defect Rates	Frequency of mistakes or need for human intervention in AI tasks.	Often hard to track consistently across different workflows.

That is not unusual for general-purpose technologies such as electricity or the internet. But financial analysts want quantitative metrics, so industries will develop them. 

Industry	Metric Category	Specific Proxy Metric
Manufacturing	Operational Efficiency	Reduction in equipment downtime (via predictive maintenance).
Healthcare	Clinical Efficiency	Time reduction for diagnostic tasks or patient documentation.
Retail	Revenue & Customer	Increase in conversion rates or uplift in average order value.
Finance	Risk & Compliance	Reduction in fraud false-positive rates or manual audit hours.
Cross-Industry	Strategic Value	Revenue generated from AI-enabled new product lines.
Cross-Industry	Human Capital	Shift in employee time from routine tasks to high-margin work.

All of these metrics can be imprecise. It can be hard to isolate AI impact from all other organizational processes, for example.